How do you read a DMARC check result (p, pct, rua, ruf, alignment)?
AdminKit's DMARC check reads your domain's _dmarc TXT record tag by tag: p is the headline policy (none, quarantine or reject) for mail failing SPF or DKIM alignment, sp sets a separate policy for subdomains, pct controls what fraction of failing mail it applies to, rua/ruf are reporting addresses, and adkim/aspf set how strictly the From: domain must match.
By AdminKit team, Netbase JSC · Updated
Key takeaways
- A DMARC result is a set of tags, not one verdict. Read
p(andsp) for what happens to failing mail,rua/ruffor who hears about it, andadkim/aspffor how strict the match is. p=noneis monitoring only. Failing mail is delivered as normal; onlyp=quarantineorp=rejectasks receivers to act.- A record with no
ruagives you a policy but no reports. Add one before you tighten anything. - A
ruaaddress at another domain only works if that domain publishes a_report._dmarcauthorization record. - RFC 7489 was obsoleted in May 2026 by RFC 9989, 9990 and 9991. Its tags still parse, but
pct,rfandriare now marked historic.
What does AdminKit’s DMARC check show you?
AdminKit’s DMARC check fetches the TXT record published at _dmarc.<yourdomain> and reads it in two ways. Both follow the current specification, RFC 9989, and still read the tags of the original 2015 specification, RFC 7489 (its IETF Datatracker record is the canonical entry, and the dmarc.org overview is a plain-language companion). The form on the tool page shows a quick parse with one plain-language line per point. AdminKit’s Check API (POST /v1/check/mail), which also feeds the AI explanation on the tool page, runs a separate checker that applies the same rules and returns findings with stable ids and severities. This guide is part of AdminKit’s email authentication cluster, alongside the related SPF and DKIM reading guides.
The quick result on the tool page has four parts:
| Part of the result | What it contains |
|---|---|
| Record | The raw TXT string found at _dmarc.<domain> |
| Tag grid | Each tag and value; p, sp, pct, rua, ruf, adkim and aspf get plain labels, any other tag is listed under its raw name |
| Findings | One plain-language line per point: the policy (including a missing p read as none), sp if set, invalid np, psd or t values, test mode, pct below 100 under quarantine or reject, historic tags, rua present or missing, ruf missing, both alignment modes, and unauthorized external report addresses |
| Suggested fixed record | Your valid p (none if p is missing but rua is valid, and also none, monitoring first, if there is no record or no valid policy), your rua (or a placeholder at your domain), your valid adkim and aspf (or r when absent or invalid), and no pct: the same form the Check API suggests |
The quick parse applies the same rules as the Check API findings listed below, without ids or severities, except where this guide notes a difference. It also adds plain informational lines that have no Check API finding (for example rua present, ruf missing and each alignment mode), and it has no organizational-domain fallback.
Both the quick parse and the Check API only count TXT strings that follow the RFC 9989 §4.8 grammar: v must be the very first tag, with nothing before it, the value DMARC1 in capitals, and then ; or the end of the record. So v=dmarc1, a leading space and v=DMARC1 p=reject (no ;) are all discarded, as RFC 9989 §4.7 says receivers do. If none is found the check reports no record; if two or more are found it reports that receivers must ignore them, which follows RFC 7489 §6.6.3 and RFC 9989 §4.10 (“they are all discarded”). In the Check API, a TXT that carries v=DMARC1 but fails that test is reported as dmarc.version_not_first. The Check API reports no record as dmarc.missing; for several records at one name it reports dmarc.multiple_records, discards them and continues the tree walk. All three are high severity.
The quick parse queries the exact name you enter. Receivers go one step further: if a subdomain has no record, they look higher up (the organizational domain in RFC 7489 §6.6.3, a DNS tree walk in RFC 9989 §4.10). A “no record” result for shop.example.com in the quick parse may therefore still be covered by example.com’s sp policy, so check the parent domain too. The Check API does this for you: it runs the RFC 9989 tree walk (at most eight _dmarc lookups), and with no record at the name you enter it applies the inherited record’s sp, or its np when the name does not exist (NXDOMAIN), and says so in dmarc.organizational_fallback_used.
Which findings does the Check API report?
These are the finding ids AdminKit’s RFC 9989 checker returns. A high finding makes the result fail, except dmarc.missing, which returns a warning; medium or low makes it a warning; info never changes the status. A DNS lookup failure returns an error with dmarc.resolver_error (medium).
| Finding id | Severity | When |
|---|---|---|
dmarc.missing | high | No DMARC record at the name or anywhere on its DNS tree walk |
dmarc.multiple_records | high | More than one DMARC record at a name; they are discarded and the tree walk continues |
dmarc.version_not_first | high | A TXT at the name carries v=DMARC1 but does not start with it, so it is discarded |
dmarc.invalid_policy_tag | high or medium | High: p invalid or missing and no valid rua URI, so receivers apply no DMARC processing. Medium: p invalid but rua valid, evaluated as p=none (RFC 9989 §4.10.1) |
dmarc.policy_not_applied | high | sp or np invalid and no valid rua URI, so receivers apply no DMARC processing (RFC 9989 §4.10.1) |
dmarc.missing_policy_tag | medium | No p but a valid rua URI: evaluated as p=none (RFC 9989 §4.10.1) |
dmarc.policy_none | medium | Policy in effect is none |
dmarc.policy_quarantine | low | Policy in effect is quarantine |
dmarc.no_rua | medium | No rua address |
dmarc.external_destination_unverified | medium | A rua/ruf domain has not published the _report._dmarc authorization |
dmarc.test_mode | low | t=y under quarantine or reject |
dmarc.partial_enforcement | low | pct below 100 under quarantine or reject |
dmarc.invalid_pct | medium | pct outside 0-100 |
dmarc.invalid_sp, dmarc.invalid_np, dmarc.invalid_psd, dmarc.invalid_t | low | Value outside the RFC 9989 §4.7 set |
dmarc.invalid_alignment_mode | low | adkim or aspf not r or s |
dmarc.invalid_report_uri | low | A rua/ruf URI that is not mailto: |
dmarc.historic_tags | info | pct, rf or ri present (historic in RFC 9989 §9.3) |
dmarc.organizational_fallback_used | info | Policy taken from the organizational (or public suffix) domain’s record, with sp or np applied |
dmarc.psl_tree_walk_differ | info | A Public Suffix List (RFC 7489) receiver would pick a different organizational domain or record |
dmarc.tree_walk_incomplete | info | A tree walk lookup above the name failed, but a record was still found |
dmarc.invalid_fo | info | Unrecognised fo value |
dmarc.report_destination_below_psd | info | A report address sits below your domain, but your record has psd=y, so a receiver may treat that address as a separate organizational domain |
dmarc.report_destinations_capped | info | More than 10 external report domains; only the first 10 were looked up |
For a domain with no record, the Check API suggests v=DMARC1; p=none; rua=mailto:dmarc-reports@<domain>; adkim=r; aspf=r, with no pct tag: monitoring first, as in the action table below. For a record with no valid policy it suggests the same, but keeps a valid adkim=s or aspf=s the record already has.
How do you read an example record?
The following is an illustrative record for example.com, annotated tag by tag. It is not a recommendation for every domain:
_dmarc.example.com. TXT "v=DMARC1; p=quarantine; sp=reject; rua=mailto:dmarc@example.com; ruf=mailto:dmarc-forensic@example.com; fo=1; adkim=r; aspf=s"
v=DMARC1 version; must come first
p=quarantine failing mail from example.com is treated as suspicious
sp=reject failing mail from any subdomain is rejected
rua=mailto:dmarc@example.com aggregate reports go here
ruf=mailto:dmarc-forensic@... failure reports go here, if a receiver sends them
fo=1 ask for a failure report if any mechanism fails to align
adkim=r DKIM: relaxed alignment (the default)
aspf=s SPF: the envelope domain must match From: exactly
AdminKit’s quick parse would list every tag in its grid, with v and fo under their raw names. It would give findings for p, sp, rua, adkim and aspf; ruf gets a finding only when it is missing. The Check API would return a single finding, dmarc.policy_quarantine (low), so the result is a warning rather than a pass.
What does each DMARC tag mean?
The first eleven tags below were defined in RFC 7489 §6.3; np, psd and t were added by RFC 9989 §4.7. The status column comes from the DMARC Tags registry update in RFC 9989 §9.3.
| Tag | Meaning | Default | Status per RFC 9989 |
|---|---|---|---|
v | Version; must be DMARC1 and the first tag | required | active |
p | Policy for the domain: none, quarantine, reject | required | active |
sp | Policy for subdomains; same values as p | inherits p | active |
pct | Percentage of failing mail the policy applies to | 100 | historic |
rua | Where aggregate reports go (DMARC URIs, usually mailto:) | none: no reports | active |
ruf | Where message-specific failure reports go | none: no reports | active |
adkim | DKIM alignment: r relaxed, s strict | r | active |
aspf | SPF alignment: r relaxed, s strict | r | active |
fo | When to send failure reports: 0, 1, d, s; ignored without ruf | 0 | active |
rf | Failure report format; only afrf defined | afrf | historic |
ri | Requested seconds between aggregate reports | 86400 | historic |
np | Policy for non-existent subdomains; same values as p | inherits sp, then p | active |
psd | Whether the record is published by a public suffix domain: y, n, u | u | active |
t | Test mode: y asks receivers to apply one level below the policy | n | active |
Unknown tags must be ignored (§6.3), so a typo in a tag name silently drops that setting rather than breaking the record.
What does each policy do to failing mail?
The p tag is the core of the record. RFC 7489 §6.3 is explicit that p=none carries no enforcement effect at all. A domain that has only ever published p=none is not protected from spoofing in any practical sense, even though a DMARC record technically exists.
| Policy | RFC 7489 §6.3 wording | In practice | Quick parse | Check API |
|---|---|---|---|---|
p=none | “no specific action be taken regarding delivery” | Failing mail is delivered; you get reports | Monitoring-only | dmarc.policy_none (medium) |
p=quarantine | Treat failing mail “as suspicious” | Spam folder, extra scrutiny, or a warning flag | Sent to spam/junk rather than inbox | dmarc.policy_quarantine (low) |
p=reject | Reject failing mail; rejection SHOULD happen during the SMTP transaction | The sender gets a bounce | Strictest enforced policy | no policy finding |
RFC 9989 §4.7 keeps the three values; for none it now says “the Domain Owner offers no expression of preference”.
Reading p=quarantine or p=reject is the signal that enforcement is actually happening.
How does sp change the result for subdomains?
sp lets a domain apply a different policy to subdomains than to the domain itself. If absent, subdomains inherit whatever p specifies (§6.3). This is useful in a specific, common situation: an organization’s main domain sends business-critical mail and needs a cautious, staged rollout to p=reject, while its many unused or rarely-used subdomains have essentially no legitimate mail traffic to break. Publishing sp=reject locks down mail spoofed from random subdomains well before the main domain’s policy catches up.
RFC 7489 notes that sp is ignored in records published on subdomains, due to how policy discovery works (§6.6.3). DMARC for subdomains explains which record a receiver applies to a subdomain and when one needs its own _dmarc record.
The Check API flags an sp value other than none, quarantine or reject as dmarc.invalid_sp (low), and an invalid np as dmarc.invalid_np (low). RFC 9989 §4.10.1 says receivers then act as if a record containing p=none was retrieved when rua holds a valid URI, and apply no DMARC processing otherwise. The Check API evaluates the record the same way: with a valid rua, the policy in effect is none for the domain and its subdomains, even if p=reject; without one, the result fails with dmarc.policy_not_applied (high). The quick parse on the tool page only flags the invalid value and keeps showing the p policy.
How much enforcement does pct leave in place?
pct applies the policy to only a percentage of failing mail. Under RFC 7489 §6.6.4, mail not selected under p=reject should be treated as quarantine, and mail not selected under p=quarantine gets the receiver’s normal classification. A result showing p=reject; pct=25 therefore means only a quarter of failing mail is rejected; the rest is quarantined.
RFC 9989 Appendix A.6 explains why pct is now historic: in practice it “was usually not accurately applied” for values other than 0 and 100. Its replacement is the t tag: t=y asks receivers to apply one level below the stated policy while you test. Some vendor guides still use the older mechanism: Google Workspace’s DMARC setup guide recommends including pct while rolling out DMARC, starting with a small percentage and raising it until it reaches 100.
The Check API lists any pct (and rf or ri) as dmarc.historic_tags, an info finding that does not change the result. If pct is below 100 under quarantine or reject, it also adds dmarc.partial_enforcement (low), whose fix is to remove pct once you trust the policy, since the default is full enforcement. t=y under quarantine or reject gets dmarc.test_mode (low), naming the weaker policy receivers are asked to apply; under none, t has no effect and gets no finding. The quick parse on the tool page gives the same lines, without ids or severities.
Where do rua and ruf reports go?
rua is the address that receives aggregate reports: XML summaries of pass/fail volume by sending source, sent daily by default (ri=86400). It is the practical foundation for knowing whether a stricter policy is safe to adopt; see the DMARC aggregate report (rua) glossary entry. ruf is a separate, optional address for per-message failure reports, and RFC 7489 §6.3 defines the two independently.
Expect few ruf reports. RFC 9989 §10.2 states that “most Mail Receivers refuse to send them out of privacy concerns.” A result with a policy but no rua address means the domain has enforcement without the visibility needed to tune it, which is worth fixing even on an already-enforced domain.
What if your rua points to another domain?
Receivers only send reports to a different organizational domain if that domain agrees. RFC 7489 §7.1, now carried into RFC 9990 §4 for aggregate reports and RFC 9991 §5 for failure reports, has receivers build the name <your domain>._report._dmarc.<report host> and look for a TXT record there starting with v=DMARC1. Without it, the address must be ignored.
If example.com publishes rua=mailto:reports@example.net, the illustrative authorization record that example.net would publish looks like this:
example.com._report._dmarc.example.net. TXT "v=DMARC1"
A report receiver willing to accept reports for any domain can publish a wildcard, *._report._dmarc.example.net. AdminKit’s Check API looks up this authorization record once for each distinct mailto: domain in rua or ruf that is not your domain, the organizational domain its tree walk found, or a subdomain of that organizational domain (RFC 9990 §4 only asks for the lookup when the organizational domains differ), up to 10 such domains per check (beyond that it adds dmarc.report_destinations_capped, info, and skips the rest), and reports dmarc.external_destination_unverified (medium) if no record starting with v=DMARC1 is there (RFC 9990 §4). For an inherited policy, the lookup uses the domain the record came from, as RFC 9990 §4 asks. The quick parse on the tool page does the same lookup, but without the Public Suffix List it only skips addresses at your domain, a parent or a subdomain of it, so a sibling subdomain under the same organizational domain may be flagged as unverified.
How strict are adkim and aspf?
Alignment is the check that the domain validated by SPF or DKIM matches the domain in the visible From: address. Without it, an attacker could pass SPF for their own unrelated domain while spoofing yours in From:. RFC 7489 §3.1 defines alignment as relaxed by default: mail.example.com aligns with example.com.
Setting adkim=s or aspf=s requires an exact domain match instead. Some organizations apply that to their primary sending domain once every legitimate sender authenticates under the exact domain; DMARC alignment: relaxed vs strict covers the trade-off.
Which DMARC specification applies now?
RFC 9989, a Standards Track document published in May 2026, obsoletes RFC 7489, according to the RFC Editor’s RFC 7489 page. RFC 9990 covers aggregate reporting, and RFC 9991 covers failure reporting.
The record syntax is unchanged and v=DMARC1 stays. RFC 9989 adds np (policy for non-existent subdomains), psd (public suffix flag) and t (testing). It also treats a record with no valid p tag as p=none when rua holds a valid URI (§4.10.1). AdminKit’s Check API reads the RFC 9989 tag set: it validates np, psd and t, reports t=y as test mode, and when rua holds a valid URI it evaluates a record with a missing or invalid p, or an invalid sp or np, as p=none. It finds the organizational domain with the RFC 9989 DNS tree walk and applies an inherited record’s np when the subdomain does not exist (NXDOMAIN); the Public Suffix List is only used for dmarc.psl_tree_walk_differ, which shows where an RFC 7489 receiver would disagree. The quick parse on the tool page applies the same RFC 9989 rules to the exact name you enter, and its tag grid lists np, psd and t under their raw names. Both read tag values such as p=Reject or adkim=S case-insensitively, as RFC 9989’s ABNF defines them; rua and ruf addresses keep their case.
Which result patterns mean you should act?
| What the result shows | What it means | What to do |
|---|---|---|
| No record | Receivers get no instruction and send no reports | Publish p=none with a rua you read |
| More than one record | Receivers discard them all (RFC 9989 §4.10) | Merge into a single record |
Missing p | RFC 7489 §6.6.3 and RFC 9989 §4.10.1 fall back to p=none if rua is valid; the Check API (dmarc.missing_policy_tag) and the quick parse do the same, and report the record as invalid without a valid rua | Add an explicit p |
p=none with rua | Monitoring is working | Review reports, then move to quarantine |
p=reject, no rua | Enforced but blind | Add rua |
pct below 100 | Partial enforcement with a historic tag | Remove pct; use t=y per RFC 9989 to test |
t=y | Test mode: receivers apply one level below the policy | Remove t=y once reports look clean |
External rua domain | Reports need authorization (RFC 9990 §4) | Check the _report._dmarc record exists |
How do you put a result together?
A fully read DMARC result answers three questions at once: what happens to failing mail right now (p, sp, pct), who finds out about failures (rua, ruf), and how strict the match has to be (adkim, aspf). Run AdminKit’s DMARC check against your own domain to see these read together in plain language, or audit SPF and DMARC with an AI agent to get the same read from a chat prompt via AdminKit’s MCP server.
See SPF, DKIM and DMARC explained for how the record fits with SPF and DKIM. To build the next record, the DMARC record generator writes the TXT string in your browser. Once reports arrive, the DMARC report analyzer reads one aggregate file. If you’re ready to move past p=none, reading DMARC aggregate reports and rolling out to p=reject safely covers the staged path from monitoring to full enforcement.
This guide is general information, not a security audit of your specific network — see our disclaimer for what AdminKit's guidance does and does not commit to.
Frequently asked questions
My check shows p=none — is my domain protected from spoofing?
No. RFC 9989 §4.7, the current DMARC specification, says that with p=none "the Domain Owner offers no expression of preference" (the original RFC 7489 §6.3 asked that "no specific action" be taken) — it only requests reporting, not enforcement. A domain at p=none is still exactly as spoofable as one with no DMARC record at all; the only thing p=none adds is visibility into failures through rua reports, which is meant as step one before moving to a stricter policy, not the end state.
What's the difference between p and sp in the result?
p sets the policy for the domain itself (mail claiming to be from example.com). sp optionally sets a separate policy specifically for subdomains (mail from mail.example.com or any other subdomain) — if sp is absent, subdomains inherit the p value. This matters because many domains send little or no mail directly from subdomains, making sp=reject a comparatively low-risk way to lock down subdomain spoofing even before the main domain reaches p=reject.
The result shows pct=50 under p=reject — what does that actually do?
Under RFC 7489 §6.6.4 it applies the reject policy to only 50% of mail that fails, selected statistically by the receiver, while the rest is treated as if the next-weaker policy (quarantine) applied. It was designed as a gradual-rollout mechanism. RFC 9989, which obsoleted RFC 7489 in May 2026, marks pct as historic because receivers applied values other than 0 and 100 inconsistently, and replaces its useful part with the t (testing) tag.
What does strict versus relaxed alignment (adkim/aspf) mean in the result?
Relaxed alignment (the default, adkim=r or aspf=r) accepts a match between the visible From: domain and the authenticating domain if one is a subdomain of the other — mail.example.com aligns with example.com. Strict alignment (adkim=s or aspf=s) requires an exact match with no subdomain exception. Most domains use the relaxed default; strict alignment is a deliberate tightening some organizations choose for their primary domain.
The check shows no DMARC record at all — what happens to spoofed mail then?
With no DMARC record, receivers have no explicit instruction from the domain owner about what to do with mail that fails SPF or DKIM alignment, and no reporting address to flag failures to — each receiver falls back to its own independent spam-filtering judgment, which varies and isn't something the domain owner can see or influence. Publishing even a p=none record with a monitored rua address is a meaningful first step, since it costs nothing in mail delivery and starts the reporting feedback loop RFC 7489 is built around.
My rua address is at another company's domain. Will reports arrive?
Only if that domain authorizes it. RFC 9990 §4 (aggregate reports) and RFC 9991 §5 (failure reports), carrying forward RFC 7489 §7.1, have receivers look up a TXT record at <your domain>._report._dmarc.<their domain> when the report host's organizational domain differs from yours. If no record starting with v=DMARC1 is found there, receivers must ignore that address. AdminKit's Check API looks this record up once for each external mailto: domain, up to 10 domains per check, and reports dmarc.external_destination_unverified when it is missing.
Sources
- Set up DMARC | Security & data protection | Google Workspace Help
- RFC 7489 — Domain-based Message Authentication, Reporting and Conformance (DMARC)
- dmarc.org — DMARC overview
- IETF Datatracker — RFC 7489, DMARC (canonical document record)
- RFC Editor — RFC 7489 info page (obsoleted by RFC 9989, 9990, 9991)
- RFC 9989 — Domain-Based Message Authentication, Reporting, and Conformance (DMARC)
- RFC 9990 — DMARC Aggregate Reporting
- RFC 9991 — DMARC Failure Reporting
Want AdminKit to watch this for you?
Free checks tell you the state today. Protective DNS and monitoring tell you when it changes.
Talk to an expert
Huy Nguyen (David), Founder & CEO, Netbase JSC
Ask whether this fix applies to your setup, or get a second pair of eyes on your DNS and mail records.